The Role of the Peripheral and Central Nervous Systems in Rotator Cuff Disease

Abstract

Rotator cuff (RC) disease is an extremely common condition associated with shoulder pain, reduced functional capacities and impaired quality of life. It primarily involves alterations in tendon health and mechanical properties that can ultimately lead to tendon failure. RC tendon tears induce progressive muscular changes that negatively impact surgical reparability of the RC tendons and clinical outcomes. At the same time, a significant base of clinical data suggests a relatively weak relationship between RC integrity and clinical presentation, emphasizing the multifactorial aspects of RC disease. This review aims to summarize the potential contribution of peripheral, spinal and supraspinal neural factors that may: (i) exacerbate structural and functional muscle changes induced by tendon tear, (ii) compromise the reversal of these changes during surgery and rehabilitation, (iii) contribute to pain generation and persistence of pain, iv) impair shoulder function through reduced proprioception, kinematics and muscle recruitment, and iv) help to explain interindividual differences and response to treatment. Given the current clinical and scientific interest in peripheral nerve injury in the context of RC disease and surgery, we carefully reviewed this body of literature with a particular emphasis for suprascapular neuropathy that has generated a large number of studies in the past decade. Within this process, we highlight the gaps in current knowledge and suggest research avenues for scientists and clinicians.

Introduction

The human shoulder complex exhibits a unique anatomical design to allow a wide range of motion at various speed and force levels. The shoulder joint complex has an unstable bony configuration secured by connective tissues and dynamic stabilizers (rotator cuff muscles) controlled by a sophisticated neuromuscular system^{156; 160}. As a consequence, shoulder structures, particularly rotator cuff (RC) tendons, are prone to various injuries and degenerative disorders^{19; 120}. RC tendon tears are common in the general population^{103; 122} and can lead to shoulder pain, impaired functional capacities, and reduced quality of life^{87; 163}.

RC tendon tears are not necessarily associated with pain or patient-reported loss of shoulder function^{90; 163; 164}, however, asymptomatic patients may develop symptoms in a relatively short period of time¹⁰⁶. Symptomatic patients usually undergo surgery when nonoperative and pharmacological options have been exhausted^{111; 129}. Surgical management decisions are mainly driven by patients’ pain, disability, and functional requirements rather than the severity of local-tissue damage¹⁵. In the short-term, nonoperative treatment may be effective in a fraction of patients^{35; 50; 75} but tissue damage and symptoms may progress over time^{90; 106; 163}, further limiting surgery and rehabilitation^{78; 91; 95; 97}. RC tendon repair is not universally successful, ~25% of repairs fail to reestablish the integrity of the rotator cuff⁹⁷ (up to 70% in massively retracted tears³⁶) and patient-reported improvements are limited^{78; 97}. Pre-operative factors such as age, chronicity, and severity of muscle-tendon unit impairments have been repeatedly associated with higher retear-rates and poorer clinical outcomes^{78; 97}. Paradoxically, two recent meta-analyses^{97; 129} suggested that patients with intact repairs might not have significant differences in symptom improvement compared to patients with recurrent tears. Another major concern is that muscle impairments do not seem to reverse, even when repair is intact and function improved at follow-up²⁶.

During the past decades, RC disease has been extensively investigated within the framework of tendon pathophysiology, tendon-to-bone healing, and muscular changes following tendon tear^30;71. A smaller set of studies have investigated how peripheral, spinal, and central neural factors are likely to contribute to muscle-tendon unit changes, impaired shoulder function, and responses to treatment. Expanding our knowledge, or at least considering the potential involvement of both peripheral and central nervous system is critical to improve our understanding of RC disease and our ability to appropriately intervene along the continuum of RC injury processes. Therefore, this review aims to scrutinize and highlight the gaps in current knowledge regarding the nervous system that may be altered in patients with RC disease from the peripheral receptors to the brain and from the brain to the neuromuscular junction. We summarized how these factors may (i) exacerbate structural and functional muscle changes induced by tendon tear, (ii) compromise the reversal of these changes during surgery and rehabilitation, (iii) contribute to pain generation and persistence iv) impair shoulder function by impairing shoulder proprioception, kinematics and muscle recruitment, iv) contribute to explain interindividual differences in symptoms and response to treatment. Given the current and lively interest for peripheral nerves injuries in the context of RC disease and surgery, we carefully reviewed this body of literature with a particular emphasis for suprascapular nerve injury that has generated a large number of studies in the past decade. Within this process, we highlighted the gaps in current knowledge and suggested research avenues for scientists and clinicians.

Proprioceptors and Related Spinal Reflexes

Shoulder movements and positional changes induce a deformation of tissues surrounding joints, including skin, muscles, tendons, fascia, joint capsules, and ligaments^{24; 27; 47; 121; 143; 155}. All these tissues are innervated by mechanically sensitive receptors termed proprioceptors that relay information to the central nervous system regarding movement, position, and forces exerted on shoulder structures (e.g. muscle spindles, Golgi tendon organs, Ruffini endings Pacinian and Meissner corpuscles). The distribution and the function of proprioceptors in shoulder joints and soft tissue have been investigated in both animal and human studies ^{40; 51; 138; 140; 143; 146; 155}).

Glenohumeral joint and ligaments receptors probably play a minor role in shoulder proprioception¹²¹ as illustrated by the small proprioceptive deficit observed after shoulder arthroplasty²¹. However they may act as limit detectors triggering protective and synergistic reflex muscle activity during movement^{27; 46; 64; 140; 148; 157}. In RC muscles and tendons, a large concentration of muscles spindles and Golgi tendon organs have been demonstrated in rabbits and rats^{3; 22; 104; 165} but no human data exist. Current theory suggests that muscle spindles are the most important proprioceptors, especially during movement¹²¹. They also play a critical role in regulating muscle contraction via spinal reflexes, that are essential for joint stability and accurate motor control¹⁰⁰. Golgi tendon organs are equally important proprioceptors, signaling information about force and mass and are also involved in the regulation of muscle contraction¹²¹.

The effect of tendon disruption on muscle spindles and Golgi tendon organs has been studied in a limited number of animal experiments concerning hind limb muscles only. Following tenotomy, muscle shortening and changes in the surrounding extrafusal tissue modify the morphology of muscle spindles that become slack and distorted¹⁶⁸. In the chronically tenotomized muscle, atrophy of intrafusal fibers, degeneration of supplying axons and fibrotic thickening of the capsule have been reported^{67; 94} Functionally, acute tenotomy decreases muscle spindle discharge^{56; 159; 168} but interestingly, responsiveness of muscle spindles from the chronically tenotomized muscle has been shown to increase^{56; 57; 168}. Shortening of intrafusal fibers, increased preliminary stretch caused by kinking of intrafusal fibers, change in passive mechanical properties or increased sensitivity of spindles have been subsequently proposed as potential explanations for this phenomenon. These increases in muscle/tendon afferent outflow have also been suggested to result from nonproprioceptive discharge^{57; 77}. Increase in the amplitude of the monosynaptic reflex has also been repeatedly observed in the chronically tenotomized muscle^{10; 61; 74; 159}, suggesting adaptive changes in motoneurons excitability consistent with the decrease in muscle mechanical loading⁹⁸. In the Golgi tendon organs, tenotomy also induces morphological changes, but the physiological consequences remain to be investigated⁶⁷. To the best of our knowledge, only one study related to proprioceptors function in RC tendon tear have been conducted and reported that experimentally-induced inflammation within rabbit RC sensitize and increase the firing of mechanical receptors¹⁶⁵.

Based on the findings of the aforementioned studies, it is reasonable to speculate that RC tendon tear is associated with structural and functional alterations of proprioceptors. Either reduced or inconsistent proprioceptive information from the injured muscle-tendon unit and altered muscle reflex activity may impair shoulder proprioception and contribute to impaired kinematics and muscle recruitment (see also section “Impact of RC Disease on Shoulder Muscle Activity”). Finally, the effects of tendon repair on the structure and the function of proprioceptors remain entirely unknown. Further experimentations are therefore required to assess the relative contribution of these mechanisms to anatomical and clinical impairments associated with RC disease.

Central Processing of Proprioceptive information

Proprioceptive information from the shoulder and more broadly from the upper limb are conveyed via the spinothalamic tracts and relayed to the somatosensory cortex where it is referred to a central body map allowing the conscious awareness of arm position and movement in space. Unconscious proprioceptive tracts (i.e. spinocerebellar tracts, projecting in the ipsilateral cerebellum) and the cervical propriospinal system are also involved the coordination movements involving multiple joints of the arm^{121; 124}.

Measurement of errors in the perceived position, movement detection latency, or ability to reproduce a determined force level can be used to globally assess shoulder proprioception^{6; 85; 107; 125; 131}. A large fraction of studies involving shoulder proprioception assessment have been conducted in patients with shoulder instability^{6; 107; 125}. In the overhead athlete with isolated infraspinatus atrophy caused by SSN compression, impaired sense of movement associated with different brain activation pattern has been reported suggesting an important contribution of RC muscle to shoulder proprioception¹³¹. Decreased sense of movement^{88; 4; 130} and a tendency to overestimate the target during force reproduction tests⁸⁹ have been reported in patients with RC tendinopathy but no data exist in patients with RC tendon tears. In conditions such as knee disorders, functional brain MRI demonstrated reduced activation of sensorimotor cortical areas and increased activation in proprioception-related brain regions, however no data exist in patients with RC disease⁶⁹. In healthy subjects, transcranial magnetic stimulation (TMS) combined with peripheral nerve stimulation has been used to assess the modulation of the propriospinal system¹²⁴ of the upper limb which is a important determinant for synergies between forearm, hand, and shoulder muscles. This system remains to be investigated in patients with RC disease.

Proprioception has been insufficiently assessed in patients with RC disease despite its recognized importance in other musculoskeletal conditions¹²¹. Therefore, further studies are required to assess proprioception in patients with RC disease and patients who have undergone RC reconstruction.

Nociceptors, Peripheral and Central Pain Processing

Nociceptors are high threshold receptors that detect signals from damaged tissue or tissue on the verge of damage. They can be found in the shoulder, skin, muscles, joints, soft-tissue, and bone^{32; 40; 41; 51; 143; 148; 149}. RC disease is associated with local-tissue damage and inflammation within the RC and surrounding structures, which release a variety of substances that sensitize nociceptors by decreasing their activation threshold (peripheral sensitization) resulting in hyperalgesia at the site of injury^{23; 34; 41}. Prolonged release of neuropeptides by nociceptive afferent fibers at the dorsal horn may sensitize nocineurons and cause long-term changes in pain processing at the spinal level and higher centers that result in pain hypersensitivity within, but also outside the original zone of injury¹⁶². As previously observed in other musculoskeletal conditions⁵³, sensory abnormalities have been observed on the injured but also on the non-injured side of patients with RC disease, illustrating the involvement of central mechanisms^{39; 48; 55}. Interestingly, patients with a RC tendon tear and signs of central sensitization have been shown to have worse clinical outcomes after surgery⁴⁸. Pain may have profound effects on motor behavior mediated at various level of the nervous system and impact on numerous motor parameters such as reflex amplitude, muscle activity, kinematics, movement planning and brain activation⁵ (see section “Shoulder Muscle Activity and Kinematics”).

Pain remains poorly characterized in patients with RC disease, but the use of existing pain assessment tools and the development of biological markers have the potential for enhancement in our understanding of pain in RC disease²⁴. Interindividual differences in the magnitude of these changes and their persistence after local-tissue damage has healed may explain differences in clinical presentation and response to therapies²⁴.

Motor Nerves and Neuromuscular Junction

The motor innervation of the RC muscles is achieved by nerves emerging from the posterior and the superior trunks of the brachial plexus, all originating from the C5–C6 cervical roots and C4 nerve root in some individuals^{2; 80; 136; 166}. The architecture and the high mobility of the shoulder complex predispose nerves to various dynamic or static compressive and/or traction injuries¹⁴⁷. Cervical radiculopathy, brachial plexopathy and peripheral nerve trunk injuries are potential comorbidities of RC tendon tear^{52; 135}. Motoneuron damage immediately reduces muscle activation and induces progressive muscle changes proportional to the severity of nerve injury¹⁴⁵. Over time, the muscle tissue can virtually disappear while connective tissue and fat accumulate⁸⁴ as recently illustrated in the human supraspinatus^{14; 79; 101}. A particular interest has been placed in the suprascapular nerve (SSN) since it innervates the most affected muscles in RC disease (i.e. supraspinatus and infraspinatus) and because it is particularly prone to entrapment^{105; 135}. SSN injury can cause shoulder weakness and pain that overlap with the signs of RC disease¹⁰⁵.

SSN injury associated with RC tendon tear, Anatomical Studies

SSN injury is possible given the surgical manipulation of previously retracted muscle(s) during RC repair procedures¹³³. In vivo studies have shown that lateral advancement during supraspinatus repair initiates a stretch of the SSN^{44; 161}. The main trunk of the SSN may be prone to damage but also its smaller branches may be injured⁴⁴. Following a similar principle, medial retraction of the supraspinatus and/or infraspinatus muscles caused by tendon tear has been suggested to place excessive traction on the SSN and to promote compressive injuries at the suprascapular and/or spinoglenoid notch. In cadavers, supraspinatus tenotomy changes the course of the SSN^{1; 93}. Various anatomical variations have also been suggested to promote suprascapular entrapment (e.g. deep and narrow shaped suprascapular notch^{60; 108; 118; 123; 150}, shape/ossification of the superior transverse scapular ligament (STSL)^{117; 119; 150}, arrangements of blood vessels^{119; 167}, configuration of the fascia securing the suprascapular nerve to the supraspinatus fossa²⁸, close relationship of the subscapularis muscle⁷). However, the incidence of these anatomical predispositions in patients with a RC tendon tear and concomitant neuropathy has never been studied. In addition, the potential occurrence of dynamic stretch/compressive strain of the SSN promoted by biomechanical and kinematic impairments in patients with RC disease should not be neglected^{20; 116}.

These anatomical studies must be acknowledged as the original incentive for investigating SSN function in RC tendon tears¹⁰⁵. However they have not addressed the question of whether these changes are physiologically relevant and whether smaller nerve branches are also likely to be insulted clinically.

Prevalence of SSN injury in patients with RC tendon tear

In patients, diagnosis of SSN injury is confirmed by electrodiagnosis that combined needle electromyography (EMG) and nerve conduction studies (NCS). Various clinical reports, retrospective studies, and prospective studies regarding the prevalence and the impact of peripheral nerve injuries before and/or after surgery have been published (see Table I for supporting material).

Table I.

Prevalence of suprascapular neuropathy in patients with rotator cuff tears before and/or after surgical repair.

		Tear etiology		Tear severity		Time of Electrodiagnosis				Prevalence of Suprascapular neuropathy
Studies	n	Trauma	Chronic	Mild/Partial	Massive/Full	Pre surgery	Post surgery	Pre-Post surgery	Time from surgery or trauma (months)	Before surgery	After surgery
Ha’eri et al. 198174	5	/	/	/	/	0	5	0	/	/	0% (0/5)
Kaplan et al. 1984101	6	6	0	5	1	5	1	1	/	100% (5/5)	100% (1/1)
Zanotti et al. 1997231	10	0	10	0	10	1	10	1	(24–36)	/	10% (1/10)
Goutallier et al. 199664	24	/	/	/	/	19	24	19	8.5	/	~21% (4/19)
Hoellrich et al. 200588	9	0	9	/	/		9		17 (6–28)	/	0% (0/9)
Vad et al. 2003211	25	8	17	17	8	25	0	0	/	8% overall (2/25)	/
Mallon et al. 2006131	8	0	8	0	9	8	4	4	6	100% (8/8)	0 % (0/6)
Costouros et al. 200726	26	19	7	0	26	26	6	6	8 (3–12)(trauma) 6 (surgery)	27% (7/26, all trauma)	0% (0/6)
Boykin et al. 201119	44	/	/	6	38	44	0	0	/	0% (partial tear) 60% (massive tear)	/
Shi et al. 2013189	60	/	/	SS (14) IS (15)	SS (46) IS (30)	60	0	0	/	29% overall (26/87) 50 % (partial SS tear) 54% (full SS tear/minor retraction) 17% (full SS tear/>5cm retraction) 20% (partial IS tear) 71% (full IS tear/minor retraction) 23 % (full SS tear/>5cm retraction)	/
Collin et al. 201424	49	24	25	0	49	49	0	0	/	2%

SS, Supraspinatus; IS, Infraspinatus; /, undocumented or unavailable information.

Following tendon repair, a low risk of iatrogenic nerve injury has been reported^{25; 49; 59; 169} but comparisons of pre- and post-surgery EMG/NCS data have not been systematically performed^{18; 42; 92; 169}. Goutallier et al.⁴² achieved such comparisons in the largest sample of patients and findings confirmed the low incidence of SSN dysfunction after RC repair previously reported. In these studies, the long time delay between surgery and electrodiagnosis may have allowed nerve recovery. Some case reports also suggested that supraspinatus repair may restore the normal course of the SSN, therefore reducing nerve strain and allowing its recovery but larger studies are required to prove this concept^{18; 92}. The large undocumented occurrence of traumatic events that could have caused direct nerve injury often limits data interpretation (see Table I).

Studies suggesting a greater prevalence of SSN motor neuropathy in patients with RC tendon tears involve important recruitment bias. In the studies of Boykin et al.¹² and Shi et al.¹³⁴, patients were sent for electrophysiological examination for persistent pain and/or severe muscle changes; Similarly, Costouros et al.¹⁸ and Mallon et al.⁹² selected patients with severe muscle atrophy and fatty infiltration. While some studies are consistent with a higher risk of SSN injury in severely versus slightly retracted tears^{13; 92}, the study of Shi et al.¹³⁴ involving a larger spectrum of RC tears severity does not support this hypothesis. These data thus call into question the concept of SSN injury as a direct consequence of muscle retraction. Prospective and carefully conducted studies indicate a rare occurrence of isolated motor SSN injury in patients with RC tendon tears, even in massive and/or traumatic RC tendon tear^{16; 153}. Within the largest patient series in this topic area¹⁶, peripheral neuropathy was found in 12% of patients and only one patient exhibited positive signs of SSN injury.

Heterogeneous and incompletely documented EMG/NCS methods are also major limitations when comparing results between these studies^{13; 92; 134; 153}. Some categorize EMG findings based upon the isolated or combined occurrence of positive EMG signs¹³⁴ while others use graded scoring based on semi-quantitative assessments of EMG abnormalities^{16; 18; 134}. Regarding NCS, some compare latencies to previously published values^{13; 16} and/or to the contralateral side^{16; 18; 153} while others compared latencies of patients with positive and negative EMG findings¹³. Severe retraction, ultrastructural muscle changes and/or non-uniform denervation may also complicate EMG/NCS in RC muscles⁸. US imaging¹¹³ and multisite EMG assessments may help overcome some of these limitations. Standardization of procedure and quantification methods¹⁷⁰ must be pursued to enhance the sensitivity of EMG. Recent progress in nerve imaging techniques^{81; 115; 144} may also allow enhancement of our ability to study peripheral nerve injuries in vivo.

Relative Contribution of Denervation to Muscle Changes associated with RC Tendon Tears: Clinical Data and Animal Models

In humans, imaging techniques cannot discriminate muscle impairments related to tendon tear or denervation when they happen simultaneously⁸. EMG/NCS is limited and an objective test such as nerve biopsy cannot be reasonably performed in patients. Consequently various animal models of RC disease and/or nerve injury have been developed to understand cellular and molecular mechanism underlying muscle changes³⁰.

In rabbits and rodents, tenotomy associated with full nerve transection has been shown to produce severe atrophy and fatty infiltration and these data are frequently used to support the role of SSN injury in human RC muscle changes^{65; 66; 72; 82; 126}. However if nerve injury occurs in humans, denervation is more likely to be incomplete with higher capacity for recovery. In rabbits, fatty infiltration has been repeatedly observed following isolated supraspinatus tenotomy^{126; 128; 152} even in absence of retraction¹⁵¹, and independent from denervation³⁸, further clouding the cause-effect relationship between nerve injury and fatty infiltration. There are many transcriptional pathways that control various aspects of the adipogenic, fibrogenic and myogenic programs⁶⁸. However, distinct pathways may be triggered by RC tenotomy or denervation as recently reported in rodents^{65; 82}. Although small animal models have a limited ability to replicate human RC disease, previously developed transgenic mice associated with tendon and/or nerve injury have great potential to further understand RC disease pathophysiology^{72; 83}. Increased availability in human tissue may also allow further investigations of muscle impairments and comparison of data obtained in animal models.

Direct Consequences of RC Tendon Tears on Nerves and Neuromuscular Junction

Studies that investigated the consequences of tendon tear on motor nerve and neuromuscular junction provide equivocal results⁶¹. These effects have been investigated in animal models of RC tendon tears in rabbit only. Signs of degenerative histological changes in the subscapular nerve after tenotomy of the subscapularis muscle have been reported¹²⁶ but characteristics of these nerve abnormalities remain unclear. Gayton et al.³⁸ reported that motor endplates were not significantly affected after tenotomy in rabbits; confirmation is required given the small sample size of this work (n=4). A critical point that has not been addressed is whether neuromuscular junctions are altered in patients with isolated RC tendon tears.

Sensory Nerves

Sensory nerve injuries have received less interest than the motor neuropathies discussed above. However the RC and surrounding structures receive sensory innervation from numerous sensory nerve branches^{29; 158} that are equally susceptible to injury. Injury within a peripheral nerve trunk induces a local inflammatory response that causes changes in afferent fibers and in the central nervous system and may lead to neurogenic pain (see section “Nociceptors and Pain Mediating Systems” and Ref.³¹ for more details). Damage to afferent fibers may also contribute to the impairment of the transduction of proprioceptive information. SSN block has demonstrated effectiveness in the management of post-operative pain⁶³ and pulsed radiofrequency modulation has been reported to provide promising long-lasting pain relief in experimental models¹⁵⁴ and in patients with shoulder pain⁶². These data highlight the important contribution of shoulder nerves in the transmission of nociceptive information in patients with RC disease, making them important targets for shoulder pain management^{63; 154}.

Shoulder Muscle Activity and Kinematics

Alterations in shoulder muscle activity and kinematics of the glenohumeral and scapulothoracic joints have been widely reported in patients with RC disease^{86; 96; 127}. One potential contributing factor may be that patients with symptomatic tears display different motor control patterns during movement compared to asymptomatic patients¹²⁷.

Kelly et al.⁷⁰ observed that symptomatic patients retain supraspinatus and infraspinatus activity despite tendon tears but are unable to activate intact deep muscles (i.e. subscapularis) as efficient co-contractors and that they may preferentially rely on periscapular muscles during elevation. These results have been partially reproduced by Cordasco et al.¹⁷ and suggest that symptomatic patients fail to develop alternative muscle activation strategies to compensate for weakened RC muscles and the resulting altered shoulder biomechanics. Importantly, they suggest that RC muscles may continue to be activated despite tendon damage. Shinozaki et al.¹³⁷ recently used positron emission tomography with fluorodeoxyglucose (FDG)^{76; 112} to assess shoulder muscle activity differences between asymptomatic and symptomatic patients. They observed increased trapezius activity and lower deltoid activity in the symptomatic group but no differences in RC muscles activity compared to asymptomatic patients. This technique appears promising but further developments are required, particularly regarding quantification.

An important issue is whether different muscle activity patterns observed in symptomatic patients are the cause or the result of pain, or both. Experimentally-induced pain has been shown to increase activity in the antagonist muscle during abduction (i.e. latissimus), probably in an attempt to limit the compression of painful subacromial structures. Similar adaptations have been observed in patients with massive RC tendon tears^{17; 142}. Masking pain may reduce these protective mechanisms and further promote local-tissue damage. Stackhouse et al.¹⁴¹ reported that pain reduced shoulder strength in external rotation in association with a decrease in voluntary activation using the twitch interpolation technique¹⁰². Sole et al.¹³⁹ also pointed out that motor adaptation to acute pain may be individual- and task-specific⁵⁸. Given the acute nature of experimentally induced pain⁵, precautions should be taken when trying to generalize these results in patients with chronic RC disease.

In patients with RC disease, pain reduction has been shown to improve glenohumeral motion and to reduce scapular contribution during arm elevation¹³². Dramatic increases of peak torque and power have also been reported⁹. Surprisingly, when assessed with isometric contractions, pain reduction has been shown to have no relevant effect on shoulder strength^{33; 114} suggesting that pain-related motor impairments may be particularly visible during movement.

These experiments observed muscle activity pattern changes under pathophysiologic and simulated conditions, however, the relative contribution of muscle-tendon unit impairments, biomechanical abnormalities, pain, impaired proprioception, and deterioration of motor control in shoulder dyskinesia and weakness remain unclear. Poor coping strategies in muscle activation patterns in response to biomechanical changes and pain may contribute to worsen local-tissue damage and pain. Interestingly, motor adaptations may also differ between individuals, in particular between symptomatic and asymptomatic patients.

Motor Cortical Changes

As in various other conditions, RC disease may induce structural and functional changes in the motor cortex that could partly explain changes in motor control and affect muscle activation. Little is known about the cortical organization of motoneurons related to proximal muscles of the arm, and even less regarding RC muscles⁹⁹. Functional MRI has been previously used but is not discriminant for motor cortical mapping of individual RC muscles⁷³. The output of the primary motor cortex (M1) can be objectively measured by motor evoked potentials (MEPs) elicited by TMS, providing direct insight on the cortical representation and the function of the corticospinal tracts⁴⁵. Mapping of the infraspinatus muscle has been recently described in healthy subjects¹¹⁰ and the same group observed positive correlation between pain chronicity and reduced M1 excitability in patients with RC disease¹⁰⁹ supporting an indirect inhibitory effect of pain on corticospinal excitability in line with current concepts⁵. However, the effects of limb disuse and other spinal/supraspinal neural factors cannot be excluded. Similarly bilateral alterations of corticospinal excitability in the deltoid and the first interosseous muscles have been reported in patients with RC tendon tears¹¹. However it should be notified that spinal motoneuron excitability must be properly assessed to verify that the change in MEPs size is not mediated at the spinal level³⁷. C3–4 propriospinal neurons may also influence the excitability of premotoneuronal sites and therefore the amplitude of MEPs^{43; 124}. Peripheral nerve stimulation associated with TMS has been recently used in healthy subjects to assess the modulation of afferent signals on M1 output⁵⁴ thus opening the possibility for its application in patients with RC disease. Further TMS studies are required to confirm the effects of RC disease on the motor cortex and to understand how these alterations may impair muscle activation, motor control, and shoulder function.

Conclusion and Perspectives

In this review, we identified a large number of neural structures and mechanisms that may contribute to pain and shoulder dysfunction in patients with RC disease. These structures and mechanisms are summarized in Figure 1. However, numerous questions remain unanswered (see Table II). Current data suggest that inflammation and muscle-tendon unit impairment disrupt proprioceptive function and reflex muscle activity. Alterations of proprioceptive afferents may impair proprioception and motor control, therefore contributing to poor muscle activation and impaired shoulder kinematics. However motor control and proprioception impairments in patients with RC diseases have been insufficiently assessed and require further investigations. Current advances in the understanding of pain pathophysiology encourage the enhancement of pain assessment and sensory abnormalities that remain poorly characterized in the clinical setting in patients with RC disease. Recent experiments suggest that the occurrence of motor nerve injury appears to be less frequent than first assumed, yet peripheral nerve dysfunction remains a non-negligible aggravating factor. Thus, this problem must be considered (perhaps with improved diagnostic tools) in clinical practice and further explored through both anatomical and physiological studies. Some data also highlight that tendon disruption, disuse, and inflammation may have a direct impact on neuromuscular junction and motoneurons but further studies are needed for confirmation. Increased availability of human tissue obtained during surgeries and animals models of RC disease will also improve our understanding of RC disease physiopathology and will help to define markers able to improve the detection of muscle denervation process. Damage inflicted to sensory nerves should not be neglected because it may contribute to the generation of pain and disrupt the afferent transduction of proprioceptive information. Evidence that RC disease induces significant motor adaptations and the important role of pain in these changes has been clearly demonstrated. However, the contribution of proprioception deficits, motor cortical changes, and modified brain activity in patients with RC disease remains to be explored. The problem of motor nervous system dysfunction is particularly relevant as the field begins to explore the mechanisms of reduced muscle force generation after reconstruction. If these problems are induced or aggravated by poor muscle activation, the nervous system impairments may need to be addressed first, and perhaps, in a way that is consistent with neurorehabilitation instead of standard musculoskeletal physical therapy. In the clinical setting, all these factors may contribute to explain why clinical presentations and responses to treatments can vary considerably between individuals despite similar peripheral tissue damage. Therefore, our final proposal is that different profiles involving different degrees of biomechanical, motor control, proprioceptive, and nociceptive impairments exist amongst patients with RC disease. The development of standardized tests achievable in the clinical setting to assess each of these aspects is necessary to provide comprehensive assessment and refine the management of these patients.

Figure 1. Potential sites for sensory and motor impairments associated with supraspinatus tendon tear.

Tendon tear, soft tissue and/or joints damage, and local inflammatory environment sensitize peripheral nociceptors (mechanical or chemical high-threshold peripheral nociceptors (e.g. Free endings) that cause pain and increase the sensitivity of central pain centers (peripheral and central sensitization, respectively). They may also induce impairments in proprioceptive outputs (Muscle spindles; Golgi tendon organ, Ruffini endings; Pacinian corpuscles) and in the central processing of proprioceptive information (proprioceptive pathways; primary sensory cortex. Motoneurons innervating both extrafusal and intrafusal muscle fibers (α- and γ– motoneurons, respectively) may equally undergo remodeling and impairments. Neuromuscular junction may also be altered as a result of reduced neural activity, muscle impairments, and central alterations within the motor nervous system. At the nerve level, stretch and/or compression caused by muscle retraction, mass compression, and manipulation of the previously retracted muscle or direct nerve manipulation during surgery can result in injury of both sensory and motor axons. The suprascapular nerve may be damaged at any point of its path but the suprascapular notch and the cervical roots are identified as the most common sites for injury. Nerve(s) damage can further increase pain, limit the afferent transduction of proprioceptive information, and aggravate muscle changes.

Table II.

Suggested deleterious nervous consequences of rotator cuff (RC) disease in studies cited in the current review.

Structures/mechanisms	Consequences	Human RC studies	Animal RC studies	Human or Animal non-RC studies
Proprioceptors, Afferences and Related Spinal Reflexes	Structural/Functional impairments of proprioceptors		226	7; 32; 85; 86; 98; 99; 112; 128; 135; 146; 153; 217; 226; 229.
	↑ Motoneuron excitability			16; 91; 108; 217
Central Processing of Proprioceptive Afferences	↓ Sense of position	180; 182
	↓ Sense of movement	126
	↓ Sense of force	8; 127
	Modified brain activity			102
Nociceptors, Peripheral and Central Pain Processing	Peripheral sensitization	51; 62; 63; 207	204
	Central sensitization	56; 73; 83
Motor Nerves	± Injury:
	Iatrogenic	74 101 231 231 160] 66; 222
	Direct consequence of RC disease	101; 131; 211 19; 24; 131 4; 14; 134	96; 97; 106; 118; 119; 176; 178; 209; 210
Neuromuscular junction	↓= Acetyl choline receptors ↓ Cholinergic/non-cholinergic muscle stimulation		55; 132; 176	92
Sensory Nerves	± Injury	75; 158		85; 135; 217; 229
Shoulder Muscle Activity and Kinematics	Modifications of muscle recruitment and kinematics	103; 123; 137; 177 25; 185; 192; 194; 197; 198 144
	↓ Voluntary activation ↑= Strength with pain reduction	15; 47; 162
Cortical changes	↓ Corticospinal excitability	17; 155

Numbers refer to references; ↓, decreased; ↑, increased; =, unchanged; review articles excluded.